Factors associated between behavior of administrating or recommending mumps vaccine and primary care physicians’ knowledge about vaccination: A nationwide cross‐sectional study in Japan

Abstract Background In Japan, the mumps‐containing vaccine was withdrawn from routine vaccination in 1993, and it became a voluntary vaccination. This study aimed to evaluate the association between the physicians’ knowledge about vaccinations and the administration or recommendation of the mumps vaccine. Methods We conducted a nationwide cross‐sectional study targeting primary care physicians (PCPs) in Japan. We used a web‐based self‐administered questionnaire by Preventive Medicine and Health Promotion Committee Vaccine Team, the Japan Primary Care Association (JPCA), from March to June in 2019. The outcome of the study was the association between PCPs’ knowledge about vaccine and the administration or recommendation of mumps vaccine. We obtained the information on background, subsidies of mumps vaccination for children from the local government, and vaccination quiz scores. We performed logistic regression analysis to estimate the odds ratios (ORs) and 95% confidence intervals (CIs). Results Among 10,470 PCPs in JPCA, 5075 were excluded. We received responses from 1084 PCPs (20.1%) and enrolled 981 participants in the analysis. PCPs with a higher score on the vaccination quizzes were significantly more likely to administrate the mumps vaccine for adults (adjusted odds ratio [AOR] 1.93, 95% CI 1.45–2.59, p < 0.001) and recommend mumps vaccine to adults than PCPs with a lower score (AOR 1.78, 95% CI 1.33–2.40, p < 0.001). Conclusions We revealed an association between the administration or recommendation of mumps vaccine and PCPs’ better vaccination knowledge.


| INTRODUC TI ON
The number of mumps infections in Japan in 2017 was 77,884; however, in the United States, where two mumps vaccinations are administered as regular immunization, it was only 6109. 1 In Japan, at least 348 people were diagnosed with deafness caused by mumps infection during January 2015 to December 2016. 2 It is important to immunize mumps vaccine not only for children but also for adults because there are reports of deafness by mumps infection in adulthood in Japan. 2 Herd immunity with better mumps vaccination coverage or rate is required to reduce the frequency of deafness from mumps infection.
We require 85-90% herd immunity against mumps infection. 3 In Japan, the measles-mumps-rubella vaccine was regulated as a routine vaccine in 1989, just as in overseas countries. This was, however, discontinued in 1993 owing to an unexpectedly high prevalence of aseptic meningitis as an adverse effect of the vaccine; thus, the mumps vaccine was only provided as a single voluntary vaccination since then (measles-rubella combined vaccine remained as routine). 4 In 2015, the vaccination rate remained only 30-40%. 5 Therefore, the authors considered that the vaccination with the local government's subsidy could effectively improve the vaccination rate by reducing the economic burden. According to a national survey in 2020, 42% (full subsidy 115 and partial subsidy 361, out of 1125) of local governments subsidized the mumps vaccination. 6 We searched mumps vaccination records of medical college students from 2008 to 2009 and found that only 58% of students (552 out of 949) received the mumps vaccine. In addition, perinatal history, past medical history, presence of siblings, and history of going abroad were not associated with mumps virus antibody positivity. 7 It is necessary for physicians to change their behavior to vaccination. Six main factors specific to physician practice regarding vaccination are guideline implementation, characteristics of practice, laws and incentives, patient characteristics/problems, social norms, and knowledge and skills. 8 After evaluating vaccine factors, we mainly focused on physician factors. Several researchers reported that recommendation by physicians affects recipients' vaccine decisions. [9][10][11][12] Physicians' appropriate knowledge about the vaccine is important for increasing vaccine administration. 13,14 Since there is no cost exemption or guidance from the government for voluntary vaccination, it is critical that it is recommended by a medical provider with appropriate knowledge. Nevertheless, the necessity of physicians' knowledge for the administration or recommendation of mumps vaccine in Japan is unknown. Thus, the study aimed to clarify the association between physicians' knowledge and the administration or recommendation of mumps vaccine.

| Study design, setting, and population
We conducted a cross-sectional study among primary care physicians (PCPs) using a web-based self-administered questionnaire by Preventive Medicine and Health Promotion Committee Vaccine Team, the Japan Primary Care Association (JPCA), which is the largest academic association for PCPs in Japan. Most JPCA physicians were internists working as PCPs at a clinic or hospital. The survey was conducted from March to June 2019 and included only JPCA members.

| Eligible Criteria
PCPs who were retired or were junior residents within 2 years after graduation from the medical school were excluded. This is because junior residents within 2 years cannot continuously work at outpatient vaccination. Further exclusion criteria included PCPs living outside Japan, those employed in a nonclinical setting, and those with missing data.

| Questionnaire
Questionnaire items were revised from previous questionnaires used by the same team. 15 We used an anonymous self-administered questionnaire and collected data on the participating PCPs' baseline characteristics such as gender, career after graduation, experience raising children, provision of daily pediatric medical service, provision of medical service at their main working facility (clinic, hospital, or other), the local government region of their main working facility, local government subsidies for mumps children vaccination, and vaccination quiz scores. We used web-based self-administered questionnaires by Preventive Medicine and Health Promotion Committee Vaccine Team, JPCA, distributed through the online mailing list of JPCA members.

| Main outcome
This study's primary outcome was an association between primary care PCPs' knowledge about vaccine and PCPs' administration of mumps vaccine for adults. PCPs were asked the following yes-no question: "Do you administer mumps vaccination for adults?" The secondary outcome of this study was an association between PCPs' knowledge about vaccine and PCPs' recommendation of mumps vaccine for adults or children. PCPs were asked a multiple-choice question with the following answers: "Actively recommend," "Recommend occasionally," "No opinion," "Not actively recommend," and "Not recommend." Answers of "Actively recommend" were defined as "recommending behavior" based on the previous research, which is a more positive behavior. 10 Furthermore, "Recommend occasionally," "No opinion," "Not actively recommend," and "Not recommend" were defined as "no recommending behavior." We obtained binary variables for the main outcome through these processes.

| Main factor
The main factor was PCPs' knowledge of vaccination, measured by a score on vaccination quizzes. The Preventive Medicine and Health Promotion Committee Vaccine Team from the JPCA prepared the quizzes by adopting the Delphi method. 16 The quizzes comprised six general vaccine questions covering Japanese vaccination affairs, including a question about mumps vaccination. The score of the quizzes was the number of correct answers to each of the six questions. We set high score as near the top 50% for acquiring points or more. We set a low score as fewer points to obtain a binary variable. We set score on the vaccination quizzes (high/low) for independent variable.

| Possible confounding factors
Possible confounding factors included experience raising children,

| Statistical analysis
We performed univariate and multivariate logistic regression analysis to estimate the odds ratios (ORs), the adjusted odds ratio(AORs), and 95% confidence intervals (CIs) using binary variables for the main outcome.
We investigated the association between PCPs' knowledge of vaccination and their administration or recommendation of mumps vaccine.
We considered submitting one independent variable as possible confounding factors per 10 events in principle. 20 In this case, we required the following variables: experience raising children, career length after graduation, possession of any specialist qualifications including primary care, information resources about vaccinations, providing medical service at their main working facility, main working region, a high or low proportion of pediatric patients in the total population, and local government subsidies for mumps vaccination of children. We then required at least 80 events for each primary or secondary outcome. We evaluated sensitivity analysis to inspect each variation for only mumps vaccination knowledge (correct or incorrect) instead of all the quizzes.
All statistical analyses were two-tailed, with significance set at p < 0.05. The analysis was performed using Stata/SE 16.2 (StataCorp LLC, College Station, TX, USA).

| Ethics
We obtained written informed consent from all participants before conducting the survey. The study protocol was approved by the Institutional Review Board at Osaka Medical College (Rin-763).

| Study flow and demographics
Among 10,470 physicians in the JPCA, 5075 who did not join the mailing list were excluded. We received responses from 1084 of 5395 PCPs (response rate: 20.1%). The respondents were from all 47 prefectures of Japan. An additional 103 participants were excluded because they lived outside Japan, they performed F I G U R E 1 Study flow nonclinical work, or they had missing data ( Figure 1). The median (interquartile range) score on vaccination quizzes was 4 (2)(3)(4)(5) points. The minimum and maximum scores were 0 and 6 points, respectively, and the mean score (standard deviation) was 3.47 (1.68) points. We set high score for 3 points or more. We set a low score as less than 3 points to obtain a binary variable. Then, participants' baseline characteristics showed that 739 (75.3%) participants were males, 358 (36.5%) have been working for 11-20 years after graduation, 420 (42.8%) were mainly working in clinics, 378 (38.6%) were working in the suburbs, and 436 (44.4%) were working in a clinical setting where the proportion of pediatric patients on the patient panel was 0-10%. (Table 1).
No other factors were associated with the recommendation of mumps vaccination for adults.

| Factors associated with recommendation of voluntary mumps vaccine for children
The classification of PCP recommendation of mumps vaccination for children showed that 731 PCPs "Actively recommend" (74.5%), 168 "Recommend occasionally" (17.1%), 57 "No opinion," (5.8%), 17 "Not actively recommend" (1.8%), and 8 "Not recommend" (0.8%) ( Note: As a result of adjusted all factors which we displayed by multivariate logistic regression analysis each other, each factor indicates an independent value for the adjustment odds ratio.

TA B L E 3 Factors associated with recommending behaviors to mumps vaccine for adults
The group that recommends vaccine against mumps (n = 327) (33.3%)

P-value
High scores in vaccination quiz (

0.23
Note: As a result of adjusted all factors which we displayed by multivariate logistic regression analysis each other, each factor indicates an independent value for the adjustment odds ratio.

0.79
Note: As a result of adjusted all factors which we displayed by multivariate logistic regression analysis each other, each factor indicates an independent value for the adjustment odds ratio.

| Sensitivity analysis
PCPs with higher scores on the vaccination quizzes were significantly more likely to administer mumps vaccination for adults than those with low scores (AOR 1.45, 95% CI 1.10-1.91, p = 0.008).
There was a positive association between voluntary mumps vaccination for adults and PCPs who worked 11-20 years after graduation, contrary to a negative association in the conventional analysis (AOR PCPs with higher scores on the vaccination quizzes were significantly more likely to recommend mumps vaccination for adults than those with low scores (AOR 2.10, 95% CI 1.59-2.78, p < 0.001).
None of factors were changed in significance with the recommendation of mumps vaccine for adults.
PCPs with higher scores on the vaccination quizzes were significantly more likely to recommend mumps vaccine for children than those with low scores (AOR 2.25, 95% CI 1.63-3.11, p < 0.001).
There was also a positive association between voluntary mumps vaccination for children and PCPs who had a higher proportion of pediatric patients (AOR 2.57, 95% CI 1.71-3.87, p < 0.001), those who works in the area where local subsidy is provided (AOR 2.44, 95% CI 1.59-3.75, p < 0.001), those who acquired information from academia (AOR 2.18, 95% CI 1.15-4.12, p = 0.02), and those who had social network service or mailing list for medical service from individual group providers (AOR 1.99, 95% CI 1.35-2.93, p = 0.001).
None of factors were changed in significance with the recommendation of mumps vaccine for children.

| DISCUSS ION
We conducted this survey among physician members of the JPCA to assess Japanese PCPs' characteristics. We found positive associations between PCPs' knowledge of vaccination and the administration or recommendation of mumps vaccine. We also found positive associations between certain information resources and the administration or recommendation of mumps vaccine.  tion. Second, we did not investigate the administration of voluntary mumps vaccine in children. Therefore, we could not symmetrically discuss the relationships between adults and children, as well as the administration or recommendation of vaccination using a two-by-two matrix. We aimed to address this logically; however, further study is required to make accurate comparisons. Third, we could not completely investigate the factors contributing to administration or recommendation of mumps vaccine owing to the cross-sectional nature of the study. Further studies are needed to confirm the proposed causality of each factor discussed in this study. Fourth, we did not adjust the effects of unknown confounding factors, which is a general limitation of observational studies. Finally, we could not obtain information from 5,075 PCPs because they did not join the mailing list.
The loss of participants led to decreased sample size. Furthermore, there might be a higher selection bias in that those JPCA members who joined the mailing list might be interested in vaccination.

| CON CLUS IONS
We revealed a significant association between PCPs' better knowledge about vaccination and the administration or recommendation mumps vaccine. Our results suggest that providing more knowledge of vaccination to PCPs may increase their likelihood to administration or recommendation of mumps vaccine and improve vaccination rate.

CO N FLI C T O F I NTE R E S T
None declared.